2006GlobecomRequires access

NXG01-2: A Weighted Fair Queuing with Optimal Rate and Delay Allocation

Tae‐Joon Kim

Open publisher page 1 citations

Abstract

The characteristic latency of each flow in the Weighted Fair Queuing (WFQ) scheduler may be over or under its delay budget allocated to the scheduler. The scheduler raises the scheduling rate to make the flow's practical latency be equal to the budget if over-budget, but otherwise it does not do anything even though there exists the excess delay resource of the difference between the budget and the characteristic latency, and, in a consequence, the resource will be wasted. In a word, legacy WFQ is non-optimal in the context of rate and delay allocation. In order to overcome this problem, this paper proposes a WFQ with optimal rate and delay allocation, called General-time Fair Queuing (GFQ).

About this research paper

What this paper is about

The characteristic latency of each flow in the Weighted Fair Queuing (WFQ) scheduler may be over or under its delay budget allocated to the scheduler. The scheduler raises the scheduling rate to make the flow's practical latency be equal to the budget if over-budget, but otherwise it does not do anything even though there exists the excess delay resource of the difference between the budget and the characteristic latency, and, in a consequence, the resource will be wasted. In a word, legacy WFQ is non-optimal in the context of rate and delay allocation. In order to overcome this problem, this paper proposes a WFQ with optimal rate and delay allocation, called General-time Fair Queuing (GFQ).

Why it matters

OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

The characteristic latency of each flow in the Weighted Fair Queuing (WFQ) scheduler may be over or under its delay budget allocated to the scheduler. The scheduler raises the scheduling rate to make the flow's practical latency be equal to the budget if over-budget, but otherwise it does not do anything even though there exists the excess delay resource of the difference between the budget and the characteristic latency, and, in a consequence, the resource will be wasted. In a word, legacy WFQ is non-optimal in the context of rate and delay allocation. In order to overcome this problem, this paper proposes a WFQ with optimal rate and delay allocation, called General-time Fair Queuing (GFQ).

Key concepts: Weighted fair queueing, Computer science, Latency (audio), Queueing theory, Scheduling (production processes), Proportionally fair, Queuing delay, Fair queuing

Related papers

Back to paper searchBrowse research topicsOriginal source
NXG01-2: A Weighted Fair Queuing with Optimal Rate and Delay Allocation — Research Paper | ScholarLens